RFR(T): 8239503: FieldLayout/OldLayoutCheck.java fails due to "RuntimeException: Misplaced int field: expected 24 to equal 12"

David Holmes david.holmes at oracle.com
Thu Feb 20 23:11:21 UTC 2020


On 21/02/2020 9:03 am, Daniel D. Daugherty wrote:
> On 2/20/20 6:02 PM, David Holmes wrote:
>> On 21/02/2020 6:38 am, Daniel D. Daugherty wrote:
>>> Greetings,
>>>
>>> We have a Tier3 failure that's causing a bit of noise in the CI.
>>>
>>> Per has determined that the test in question should not run with ZGC
>>
>> This seems a temporary workaround. Obviously there are issues with the 
>> new field layout code and interaction with ZGC that have not been 
>> accounted for. The underlying problem presumably needs to be addressed 
>> and now we won't have an open bug for that. I would have treated this 
>> like a "problem listing" and created a sub-task.
> 
> This is what Per wrote in his comment before his proposed patch:
> 
>> Using -XX:+UseCompressedClassPointers has no effect when using ZGC, so 
>> the field offset will be different. The test should just exclude ZGC. 
> 
> That doesn't sound like a temporary condition with ZGC to me.

Apologies I should have dug deeper first.

David

> Dan
> 
> 
>>
>> Cheers,
>> David
>> -----
>>
>>> and added a patch to the bug report. I'm testing that patch with a
>>> Tier3 test run right now...
>>>
>>> Here's what it looks like:
>>>
>>> $ hg diff
>>> diff -r 38fc3785784a 
>>> test/hotspot/jtreg/runtime/FieldLayout/OldLayoutCheck.java
>>> --- a/test/hotspot/jtreg/runtime/FieldLayout/OldLayoutCheck.java Thu 
>>> Feb 20 20:04:15 2020 +0000
>>> +++ b/test/hotspot/jtreg/runtime/FieldLayout/OldLayoutCheck.java Thu 
>>> Feb 20 15:20:47 2020 -0500
>>> @@ -28,7 +28,7 @@
>>>    * @library /test/lib
>>>    * @modules java.base/jdk.internal.misc
>>>    *          java.management
>>> - * @requires vm.bits == "64"
>>> + * @requires vm.bits == "64" & vm.gc != "Z"
>>>    * @run main/othervm -XX:+UseCompressedClassPointers 
>>> -XX:-UseEmptySlotsInSupers OldLayoutCheck
>>>    */
>>>
>>>
>>> Thanks, in advance, for any comments, questions or suggestions.
>>>
>>> Dan
>>>
> 


More information about the hotspot-runtime-dev mailing list